[Dale Paquette]

Dave, I appreciate your “stock” answer wrt HDV. The problem is that some of us who are serious amateurs without the wherewithal to pass off equipment costs to the client through equipment charges or to the Gov through business and depreciation expenses are kind of stuck. To advance beyond HDV requires a large capital outlay that I, for one, can’t afford anytime soon. For now, I need to find a workflow that leads to the best possible results within the limitations of what is affordable.

My video goal is to take the best footage I can, enhance them with the full range of VFXs, and present them to viewers. Until I upgrade my NLE, which I use for the burning process, I’m limited to 720p30. This is considerably lower than 1920 X 1080. However, I am often forced into handheld shots due to where I take my footage – often in places that require high climbing that do not lend themselves to bringing any more equipment than the minimal. Thus, stabilization is an important part of what I do in AE. This almost always requires scaling up some to cover edges. I’m assuming at this point that then scaling down to 720 compensates to a large degree for the softness introduced with the up scaling. So, 720 isn’t too bad from that perspective. As to how that impacts the final picture on the TV, I haven’t done conclusive tests yet as I can’t burn to 1080p30 yet. Dale

[Dale Paquette]

Sorry I seem so confused. My camera outputs HDV (which is highly compressed mpeg). That’s what I capture to my PC and that’s what I import to AE (for CC, effects, etc.)and my question relates to should I use it’s native size (1440 X 1080) for the comp or full HD. In either case, I then export it from AE as either an uncompressed avi or QT for use in my NLE (or could go with a TIFF sequence). It’s only after the final edit that I burn (directly) to Blu-ray which, if I understand it, is also mpeg (different version for Blu-ray). I know this much, the file size for a full HD comp is 34% bigger (as an avi) than for anamorphic HD and more difficult to handle. But, I can, if the end result will look any better when seen on a large LCD HD TV. If the final appearance will be the same, why not go with the smaller sizes? Thanks, Dale

[Dale Paquette]

Roland, this is my second response this morning. I did indeed read through the reference you provided and will now study it in detail. Based on my first read through, I went back to the expression and made a change – fist, I didn’t use the //Start Script notation (nor the //End Script). This isn’t mentioned in the article. But, I did have the three lines of the expression. What I did was add a semicolon after the first 2 lines per the article. As far as I can tell, this is now working. I say as far as I can tell because I guess the two stabilized layers are so close to each other that the average is basically indiscernible from its components!! So, I will try again but this time make certain that the two tracks are off from each other by enough to be very clear where the average falls. I will also try rotation this next time. Thanks again, Dale

[Dale Paquette]

Roland, I’m back on the job. Haven’t got to your reference yet but will. In the meanwhile, here’s my step-by-step interpretation of your method. To simplify, I tried to average just 2 position only stabilizations (to include rotation, I assume the expression rotation control is used instead of point control). Started with the original plus 2 copies. Stabilized position only separately for 2 of the clips. Created a null and added 2 expression point controls. For each expression point control, alt clicked the stopwatch and added an expression by pickwhipping directly to the corresponding stabilized layer’s anchor point (where stabilization keyframes are clearly evident). Then, for the original clip, alt clicked it’s anchor point stopwatch and added expression exactly as you wrote but modified for 2 stabilized layers. When I finish, I get an AE warning reading in part- Class ‘Layer” has no property or method named ‘Tracker 02 Value’ So, I’ll dig into your reference and see if I can figure it. If you aren’t getting too tired of this, I’d love to hear any further thoughts you may have. Thanks, Dale

[Dale Paquette]

Thanks, and I will. Out of action for a week or so but will get back on this soon. May I ask another question? My source footage is HDV (basically MPEG-2) which is, of course, highly compressed. As far as motion tracking is concerned, is MPEG-2 as good an any format or would I be better served by opening it in say a 1920 X 1080 square PAR and exporting it as an uncomporessed avi or the like first? Any way, I have to stop till I return home where I can acess my PC and AE. Thanks, Dale

[Dale Paquette]

RoRK, I understand.

As for averaging, I imported some footage and duplicated twice (now have 3 instances). I labeled them master and stabilized 1 and 2. I performed a “normal” stabilize position only on the last 2 and ended up with each having it’s anchor point key-framed as expected. Then, I opened a null and added two expression point controls. For each one, I pick-whipped it to one of the stabilized layer anchor points. Here’s where I get lost. Where do I add the expression? I’m thinking the null’s position (because the anchor is different for a null than a regular layer). Also, it’s not clear if I’m to copy your expression verbatim or if the references to pick-whips means to actually pick-whip. Finally, since each tracker is on a different layer, both have the same number whereas you have different numbers.

Obviously, expressions is one of my worst weak points!

I also tried going directly to the anchor point of the master layer, pick whipping it to the AP of the first stabilized layer, added +, pick-whipped to the second AP, surrounded the whole thing with ( ), and added /2. This seemed to work. I know there are advantages to going tough the null, but I’m not sure how to make it work wrt expression point controls though I think my technique would work directly with the null’s position. Are these two ways equivalent or even the same? Thanks. Dale

[Dale Paquette]

RoRK, I understand.

As for averaging, I imported footage and duplicated twice (3 copies). One I call the master and the others stabilize 1 and 2. On these last two, I do a normal stabilize position only and end up with each having its anchor point key framed as expected. Then, I open a null and add 2 expression point controls. I pickwhip each one of these to one of the stabilized layer anchor points. Now, I lose my sense of where to add the expression. Since it is a null, I’m thinking its position because the anchor point location is different for a null than a regular layer. So, I alt click the stop watch and add the script. But, for one thing, each tracker, since it’s on a different instance of the footage, has the same number whereas you have a different number. Anyway, it’s in where to apply (and how) the script.

That said, I experimented with going direct to the anchor point property of the master layer, opened an expression box, and pick-whipped it to the anchor of the first layer, +, pick-whip to the second layer anchor layer. Then I surround the whole thing with ( ) and add /2. This seems to work. I understand that there are advantages to going to a null first, but I can’t seem to figure it. Is what you suggest and what I did basically the same?

Obviously, expressions is one of my worst weak points. Dale

[Dale Paquette]

You guessed it, operator error! What I did was to try stabilization in the new comp. I fixed it by going back to the first comp and adding the second comp. After hiding the original footage, voila – a B&W hicon version. Now, opening the stabilizer leaves the hicon version in in place for tracking! BTW, I’m just desaturating and then using curves to really increase the contrast. The tracking results look good.

Now, back to the averaging. Thanks, Dale

[Dale Paquette]

Thanks for the response. Is it operator error? I’d bet on it!! That said, I’ll try your suggestion later today.

The hicon version seems to allow for better separation of highly luminescent points allowing me to make a much smaller feature box and search box thus speeding up the whole process. I’ll be busy working this some more soon as I fix my precompose problem.

So far, I have made most of the averaging part work, but seem to be having some difficulty with the expression. I’ll keep at it till I get it.

[Dale Paquette]

I’ll give it a try. Thanks. Dale